The number of patients surviving a critical illness is increasing, with an estimated 1.4 million persons over the age of 65 now surviving the ICU annually. Unfortunately, two out of three ICU survivors suffer from long-term cognitive impairment (LTCI) at 1 year after their critical illness. This cognitive impairment persists in man for at least three years following critical illness, preventing these survivors from returning to wrk and diminishing their quality of life. LTCI likely results from the interplay of host baseline facors plus acute problems associated with the patient's illness and ICU management, yet how combinations of risk factors predict development of LTCI is not well understood. Delirium, a potentially modifiable risk factor, is associated with LTCI, but whether milder forms of acute brain dysfunction (e.g., subsyndromal delirium) contribute to the development of LTCI has not been considered to date. Initial attempts at cognitive rehabilitation for suffering ICU survivors b the investigator and his team appear feasible and promising, however, neither clinicians nor researchers currently have the ability to determine who will develop this disabling problem of LTCI. Thus, a key barrier to further study of cognitive interventions for the growing population of critically ill elderly is the inability to identify patients, early in their ICU course, who are atrisk of developing LTCI. Therefore, the proposed research will derive and validate a clinical prediction tool that will accurately identify critically ill patients, early in the course of their illness, a risk of developing LTCI that persists at 1-year follow-up. The Specific Aims of this proposal are (1) To derive a clinical prediction tool for early identification of ICU patients at risk of developing log-term cognitive impairment, (2) To determine the relative value of delirium severity categorizations within the clinical prediction tool (from Specific Aim 1), and (3) To externally validate the clinical prediction tools (from Specific Aims 1 and 2) for identification of ICU patiets at risk of developing long-term cognitive impairment. This proposal integrates the two largest cohorts (including an NIA sponsored cohort) of ICU survivors with cognitive outcomes testing available to date with state of the art statistical methods for the development of clinical prediction tools to advance the science and understanding of the emerging public health problem of dementia-like LTCI for all ICU survivors and will especially benefit the rapidly growing population of elderly ICU patients. In addition to providing clinicians and patient familie with a novel lens through to view a patients' long-term cognitive prognosis, this award will allow the investigator to establish a track record of individual aging research and pave the way to a mentored career development award by allowing identification of the target population for the investigator's next phase of this unique and ground breaking research, cognitive rehabilitation of ICU survivors with LTCI.